Description
Hi there and welcome to our course; Reinforcement Learning.
Reinforcement Learning is a really thrilling and essential discipline of Machine Learning and AI. Some name it the crown jewel of AI.
On this course, we are going to cowl all the facets associated to Reinforcement Learning or RL. We are going to begin by defining the RL drawback, and evaluate it to the Supervised Learning drawback, and uncover the areas of purposes the place RL can excel. This consists of the drawback formulation, ranging from the very fundamentals to the superior utilization of Deep Learning, resulting in the period of Deep Reinforcement Learning.
In our journey, we are going to cowl, as regular, each the theoretical and sensible facets, the place we are going to learn to implement the RL algorithms and apply them to the well-known issues utilizing libraries like OpenAI Fitness center, Keras-RL, TensorFlow Brokers or TF-Brokers and Steady Baselines.
The course is split into 6 foremost sections:
1- We begin with an introduction to the RL drawback definition, primarily evaluating it to the Supervised studying drawback, and discovering the software domains and the foremost constituents of an RL drawback. We describe right here the well-known OpenAI Fitness center environments, which will probably be our playground with regards to sensible implementation of the algorithms that we find out about.
2- In the second half we focus on the foremost formulation of an RL drawback as a Markov Resolution Course of or MDP, with easy resolution to the most elementary issues utilizing Dynamic Programming.
3- After being armed with an understanding of MDP, we transfer on to discover the resolution area of the MDP drawback, and what the totally different options past DP, which incorporates model-based and model-free options. We are going to focus on this half on model-free options, and defer model-based options to the final half. On this half, we describe the Monte-Carlo and Temporal-Distinction sampling primarily based strategies, together with the well-known and essential Q-learning algorithm, and SARSA. We are going to describe the sensible utilization and implementation of Q-learning and SARSA on management tabular maze issues from OpenAI Fitness center environments.
4- To maneuver past easy tabular issues, we might want to find out about perform approximation in RL, which ends up in the mainstream RL strategies right now utilizing Deep Learning, or Deep Reinforcement Learning (DRL). We are going to describe right here the breakthrough algorithm of DeepMind that solved the Atari video games and AlphaGO, which is Deep Q-Networks or DQN. We additionally focus on how we are able to resolve Atari video games issues utilizing DQN in observe utilizing Keras-RL and TF-Brokers.
5- In the fifth half, we transfer to Superior DRL algorithms, primarily underneath a household known as Coverage primarily based strategies. We focus on right here Coverage Gradients, DDPG, Actor-Critic, A2C, A3C, TRPO and PPO strategies. We additionally focus on the essential Steady Baseline library to implement all these algorithms on totally different environments in OpenAI Fitness center, like Atari and others.
6- Lastly, we discover the model-based household of RL strategies, and importantly, differentiating model-based RL from planning, and exploring the entire spectrum of RL strategies.
Hopefully, you take pleasure in this course, and discover it helpful.
If the coupon will not be opening, disable Adblock, or attempt one other browser.